CrowdLens: AI-Based Real-Time Crowd Behaviour Analyser

CrowdLens is an AI-powered real-time crowd be-havior analyzer designed to interpret and evaluate dynamic human crowd patterns using state-of-the-art deep learning tech-niques.The system addresses the critical need for proactive crowd monitoring, especially in densely populated environments such as transportation hubs, stadiums, festivals, and public demonstrations, where unexpected crowd behavior can lead to security breaches, injuries, or fatalities.Traditional surveillance systems rely heavily on manual moni-toring, making them prone to human error, fatigue, and delayed response times.In contrast, CrowdLens leverages advanced computer vision models, including object detection (YOLOv8), crowd density estimation, and behavioral anomaly detection, to autonomously assess crowd activities from live video feeds.The system can identify events such as sudden dispersion, congestion formation, abnormal movement patterns, and aggressive interac-tions.This paper presents the design, methodology, and evaluation of the CrowdLens system.It provides a comprehensive review of related work, outlines the system architecture, discusses implementation using realtime video analytics, and evaluates performance using standard crowd datasets.Additionally, we ad-dress deployment challenges and ethical considerations, including privacy concerns and algorithmic bias.The proposed solution significantly enhances situational aware-ness and supports timely decision-making for security personnel and emergency responders.By minimizing the need for continuous human oversight and enabling automated alerts in high-risk scenarios, CrowdLens contributes to safer, smarter, and more responsive public space management.